6. ML-powered Accurate Attribute Matching

Objective:

~99% match accuracy as validated by SMEs

This is a core function to accurately identify unique entities within an enterprise e.g., unique customer entity. It is also important to provide transparency to the users and allow them to understand their existing data better as well as provide visibility into the actual matching process.

Other non-functional requirements include:

end-to-end time taken to successfully complete the matching process pipeline

volume of records able to process successfully in any single run

⁠

Attribute-level Matching

ML-powered Attribute-level Matching

Create UI to show summary of the Attribute-level Matching Process along with interesting insights

Create UIs to show visual representation of attribute-level matches between a pair of records.

The default page shows a few examples recommended by the system, but allows the users to choose pairs based on a variety of parameters (like TBD).

Priority

⁠

Data Sources for Data Quality:

BIDM

Want to print your doc?
This is not the way.

Try clicking the ⋯ next to your doc name or using a keyboard shortcut (

CtrlP

) instead.